Development of widely applicable techniques for the interpretation of computer models of metabolism by use of relatively standard methods of sensitivity analysis is proposed here. These techniques will be applicable to metabolic models of reconstituted enzymic systems or intact tissues where the kinetics of the constituent chemical reactions are represented by pesudo-steady state rate laws. A very large variety of biological systems can be modeled and analyzed within this limitation. Much of the software necessary for this project already exists, and the proposed developments are additions to a working program package. These developments will be valuable in identifying biochemical and physiological determinants of metabolic regulation. If metabolic models are ultimately to be useful in clinical applications, there must be convincing proof of their reliability. The proposed research will greatly aid in achieving this goal by providing a quantitative measure of the maximum tolerable error in a model's parameters.